***************************************************************
***		DoFile for regression on establishment rents		***
***		Last change: 20.12.2017 fo							***
***************************************************************
capture log close
log using "$log/vert_04_regress_v26.log", replace
set more off
adopath ++ $prog

* for person-level regressions
global controls = "ib0.workertype exper_estab exper_estab2 exper_estab3 exper exper2 exper3 experxfemale exper2xfemale exper3xfemale vtxexper vtxexper2 vtxexper3"

* for establishment-level regressions
global treatments = "leiharb100 leiharb100_sq freie100 freie100_sq os_panel"
global brinteract = "leiharb100xbr leiharb100_sqxbr freie100xbr freie100_sqxbr os_panelxbr sizecat2xleiharb100 sizecat3xleiharb100 sizecat4xleiharb100 sizecat5xleiharb100 sizecat6xleiharb100 sizecat7xleiharb100 sizecat2xleiharb100_sq sizecat3xleiharb100_sq sizecat4xleiharb100_sq sizecat5xleiharb100_sq sizecat6xleiharb100_sq sizecat7xleiharb100_sq sizecat2xfreie100 sizecat3xfreie100 sizecat4xfreie100 sizecat5xfreie100 sizecat6xfreie100 sizecat7xfreie100 sizecat2xfreie100_sq sizecat3xfreie100_sq sizecat4xfreie100_sq sizecat5xfreie100_sq sizecat6xfreie100_sq sizecat7xfreie100_sq sizecat2xos_panel sizecat3xos_panel sizecat4xos_panel sizecat5xos_panel sizecat6xos_panel sizecat7xos_panel"
global lage	  = "ib2.ertragslage_py ib1.gevol_cat"
global beschgesamt = "beschgesamt beschgesamt_sq lnbeschgesamt"


********************************************************************************
set matsize 1000
matrix define 	robustness = J(128 , 18, .)
matrix colnames robustness = 1_modelnr 2_Npers 3_Nest 4_bleiharb10 5_se 6_bfreie10 7_se 8_bos_panel 9_se 10_sector 11_tvcontrols 12_bleiharb10xbr 13_se 14_bfreie10xbr 15_se 16_bospanelxbr 17_se 18_brinteract

local rowcount = 0
local sectorcount = 0
foreach cond_sector in manuf other {
	local ++sectorcount
		local fecount = -1
				dis " "
				dis "`cond_sector'"

				use "$data/linked_imputed_2002-2008.dta", clear
				keep if (`cond_sector' ==1 & nocoll==1)
				gen vtxexper  = vt*exper
				gen vtxexper2 = vt*exper2
				gen vtxexper3 = vt*exper3

				qui tab sizecat, gen(sizecat)
				foreach num of numlist 1/7 {
					gen sizecat`num'xleiharb100 	= sizecat`num' * leiharb100
					gen sizecat`num'xleiharb100_sq  = sizecat`num' * leiharb100_sq
					gen sizecat`num'xfreie100	 	= sizecat`num' * freie100
					gen sizecat`num'xfreie100_sq    = sizecat`num' * freie100_sq
					gen sizecat`num'xos_panel 		= sizecat`num' * os_panel
					}

				dis "first step: regress wages on person characteristics (with occupation-education-year fixed effects and state-year fixed effects)"

					dis " "
					dis "Variablennamen im Folgenden:"
					dis "	milnwage: log. Tagesentgelt (z.T. imputiert) "
					dis "	workertype: Stellung im Beruf "
					dis "	exper_estab exper_estab2 exper_estab3: betriebsspez. Berufserfahrung (linear, quadriert, kubisch) "
					dis "	exper exper2 exper3: allgem. Berufserfahrung (linear, quadriert, kubisch) "
					dis "	xfemale: Interaktion mit dem Geschlecht "
					dis "	vtx: Interaktion mit dem Merkmal abgeschl. Berufsausbildung "
					dis " "
					summ milnwage $controls

					reghdfe milnwage $controls [pw=persweight], abs(i.yr_occ_edu_fe i.yr_bula_fe i.spell_fe) resid(resid) keepsing
						local Npers = e(N)

					* summarize residuals by establishment-year and reduce dataset to establishment-years
					keep if e(sample) == 1
					bysort estid year: egen est_resid = mean(resid)
					sort estid year persnr
					by estid year: keep if _n==1

				dis " "
				dis "second step: regression on establishment rents"
				dis " "
				dis "Variablennamen im Folgenden:"
				dis "	est_resid: Betriebs-Rents (Betriebs-Residuum aus obiger Regression auf das Entgeld)  "
				dis "	leiharb100: Anteil der Leiharbeiter an der Belegschaft (in Prozent)"
				dis "	freie100: Anteil der freien Mitarbeiter an der Belegschaft (in Prozent)"
				dis "	os_panel: Outsourcing-Ereignis"
				dis "	export_proz_py: Exportanteil am Umsatz"
				dis "	ertragslage_py: Ertragslage (ordinal)"
				dis "	gevol_cat: Entwicklung Geschäftsvolumen (kategorial)"
				dis "	beschgesamt: Beschäftigte insgesamt"
				dis "	persuche: Derzeit vakante Stellen (nein/ja)"
				dis "	brantarifv: Branchentarifvertrag (nein/ja)"
				dis "	pc_high: Anteil der Hochschulabsolventen an der Gesamtbelegschaft über dem Median in der Referenzgruppe (nein/ja)"
				dis "	betriebsrat: Betriebsrat (nein/ja)"
				dis "	*xbr: verschiedene Interaktionen mit Betriebsratstatus"
				dis "	sizecatx* verschiedene Interaktionen mit Größenklasse des Betriebs"
				dis " "

				summ est_resid leiharb100 freie100 os_panel export_proz_py i.ertragslage_py i.gevol_cat beschgesamt persuche brantarifv betriebsrat

				qui reghdfe est_resid $treatments 	$lage 	persuche 	export_proz_py 	$beschgesamt 	brantarifv 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 	persuche 	export_proz_py 	$beschgesamt 				[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 	persuche 	export_proz_py 					brantarifv	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 	persuche 	export_proz_py 								[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 	persuche 					$beschgesamt  	brantarifv	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 	persuche 					$beschgesamt 				[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 	persuche 								 	brantarifv	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 	persuche 												[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 				export_proz_py 	$beschgesamt  	brantarifv	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 				export_proz_py 	$beschgesamt  				[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 				export_proz_py 					brantarifv	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 				export_proz_py 								[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 							 	$beschgesamt  	brantarifv	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 							 	$beschgesamt 	 			[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 							 					brantarifv	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 							 								[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 	persuche 	export_proz_py 	$beschgesamt  	brantarifv	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 	persuche 	export_proz_py 	$beschgesamt  				[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 	persuche 	export_proz_py 				 	brantarifv	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 	persuche 	export_proz_py 								[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 	persuche 				 	$beschgesamt 	brantarifv 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 	persuche 				 	$beschgesamt	 			[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 	persuche 				 					brantarifv	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 	persuche 				 					 			[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 				export_proz_py 	$beschgesamt 	brantarifv	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 				export_proz_py 	$beschgesamt 				[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 				export_proz_py 				 	brantarifv	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 				export_proz_py 				 				[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 							 	$beschgesamt 	brantarifv	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 							 	$beschgesamt  				[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 							 					brantarifv	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 							 								[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
						local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
					qui lincom (10*leiharb100) + (100*leiharb100_sq)
						matrix robustness [`rowcount', 4] = r(estimate)			 	/* coef */
						matrix robustness [`rowcount', 5] = r(se) 					/* se */
					qui lincom (10*freie100) + (100*freie100_sq)
						matrix robustness [`rowcount', 6] = r(estimate)				/* coef */
						matrix robustness [`rowcount', 7] = r(se)					/* se */
						matrix robustness [`rowcount', 8] = _b[os_panel]			/* coef */
						matrix robustness [`rowcount', 9] = _se[os_panel]			/* se */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',18] = 0						/* effect heterogeneity by works council status */


			*** effect heterogeneity by works council status ***

				qui reghdfe est_resid $treatments 	$lage 	persuche 	export_proz_py 	$beschgesamt 	brantarifv	$brinteract 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
				local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */
				qui reghdfe est_resid $treatments 	$lage 	persuche 	export_proz_py 	$beschgesamt 				$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 	persuche 	export_proz_py 					brantarifv	$brinteract 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
				local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 	persuche 	export_proz_py 								$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 	persuche 					$beschgesamt  	brantarifv	$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 	persuche 					$beschgesamt 				$brinteract 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 	persuche 								 	brantarifv	$brinteract 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 	persuche 												$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 				export_proz_py 	$beschgesamt  	brantarifv	$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 				export_proz_py 	$beschgesamt  				$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 				export_proz_py 					brantarifv	$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 				export_proz_py 								$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 							 	$beschgesamt  	brantarifv	$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 							 	$beschgesamt 	 			$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 							 					brantarifv	$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 	$lage 							 								$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 	persuche 	export_proz_py 	$beschgesamt  	brantarifv	$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 	persuche 	export_proz_py 	$beschgesamt  				$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 	persuche 	export_proz_py 				 	brantarifv	$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 	persuche 	export_proz_py 								$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 	persuche 				 	$beschgesamt 	brantarifv 	$brinteract 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 	persuche 				 	$beschgesamt	 			$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 	persuche 				 					brantarifv	$brinteract 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 	persuche 				 					 			$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 				export_proz_py 	$beschgesamt 	brantarifv	$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 				export_proz_py 	$beschgesamt 				$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 				export_proz_py 				 	brantarifv	$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 				export_proz_py 				 				$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 							 	$beschgesamt 	brantarifv	$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 							 	$beschgesamt  				$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 							 					brantarifv	$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				qui reghdfe est_resid $treatments 		 							 								$brinteract	 	[pw=weight_size], abs(i.estid i.yr_ind_size_fe) vce(cluster estid) keepsing fast
					local ++ rowcount
					dis " modelnr:" `rowcount'
						matrix robustness [`rowcount', 1] = `rowcount'				/* modelnr */
						matrix robustness [`rowcount', 2] = `Npers'					/* #persons in wagereg */
						matrix robustness [`rowcount', 3] = e(K1)					/* #establishments in second step regression */
						matrix robustness [`rowcount',10] = `sectorcount'
						matrix robustness [`rowcount',11] = 0						/* time varying controls */
						matrix robustness [`rowcount',16] = _b[os_panelxbr]			/* coef */
						matrix robustness [`rowcount',17] = _se[os_panelxbr]		/* se */
					qui lincom (10*leiharb100xbr) + (100*leiharb100_sqxbr)
						matrix robustness [`rowcount',12] = r(estimate)				/* coef */
						matrix robustness [`rowcount',13] = r(se)					/* se */
					qui lincom (10*freie100xbr) + (100*freie100_sqxbr)
						matrix robustness [`rowcount',14] = r(estimate)				/* coef */
						matrix robustness [`rowcount',15] = r(se)					/* se */
						matrix robustness [`rowcount',18] = 1						/* effect heterogeneity by works council status */

				}

/*
		Die folgende Tabelle berichtet Koeffizienten von FE-Regressionen auf Betriebs-Rents.
		Die hohe Zahl an Koeffizienten (=Zeilen) ergibt sich aus der Notwendigkeit, Modelle mit allen möglichen Kombinationen an Kontrollvariablen zu berechnen
		(Vgl. Young and Holsteen (2017): Model Uncertainty and Robustness: A Computational Framework for Multimodel Analysis. Sociological Methods & Research 46: 3-40.)

		Legende:
		Npers: 	# Personen in first-stage-Regression
		Nest:	# Betriebe in second-stage-Regression
		bleiharb10, bfreie10: Koeffizient für eine Zunahme atypischer Beschäftigung um 10-Prozentpunkte.
		bos_panel:	Koeffizienten für Schließung oder Ausgliederung von Teilen des Betriebs
		se:			Standardfehler
		sector: 	1=verarb. Gewerbe und Bergbau, 2= alle anderen Branchen
		tvcontrols: Modell mit/ohne zeitveränderlichen Kontrollvariablen
*/
mat list robustness

*** END ***
log close








